Intranasal oxytocin (OT) has shown promise as an agent that can effectively improve social processing in disorders that are characterized by impairments in social behaviors, including schizophrenia. However, important gaps in our understanding of OT's effects on the brain have greatly limited its development as a therapeutic agent and made it difficult to interpret mixed research findings. One very important therapeutic function of OT may be its ability to enhance learning of social information. The proposed research will evaluate, in sequential steps, OT's effects on learning social cognitive skills in psychotic patients within a manualized training program that we have developed. The initial step (R21) is to demonstrate target engagement (i.e., that OT is getting to the brain and influencing social processes) and to establish optimal dosing of OT. Promising neurophysiological measures of target engagement for OT include EEG mu suppression while observing biological motion, and pupillary dilation during an emotion recognition task. The R21 phase will determine the dose-response curve to OT for proposed measures of target engagement (i.e., mu suppression and pupillary dilation). The proposed study will use a single administration, within-subjects, crossover design to model the dose-response to OT across a large range of doses. Subjects with schizophrenia will be randomized to either: 8, 12, 24, 36, 48, 60, 72, or 84 IU OT (6 subjects to each dose). Each subject will receive OT and placebo in administrations separated by one week in randomized order. 30 minutes following OT administration, mu suppression and pupillary response will be assessed. We will also collect data on two measures of social salience (operationally defined here as behavioral tasks of social preference) to examine their correlations with measure of target engagement. The R33 phase will determine whether OT administration 30 min before a SCST session enhances the learning of social information in the context of a social cognitive skills training (SCST) program, and it will test a possible mediator of this effect. Subjects with psychotic disorders will be randomized to one of four groups in a 2 by 2 factorial design: OT with SCST; Placebo with SCST; OT with training control condition; placebo with training control condition. Measure(s) of target engagement supported in the R21 phase, as well as measures of social preference, will be examined in an OT challenge (placebo versus OT one week apart) prior to baseline assessment. A social cognition battery will be administered at baseline, at midpoint after 6 weeks of SCST, and at completion of training at 12 weeks. The battery will include measures of social cue identification and mentalizing.